ความชุกความไม่มั่นคงทางอาหารของประชากร โดยใช้เกณฑ์การวัดประสบการณ์ความไม่มั่นคงทางอาหาร

44 Troubleshooting The tables below can be used to summarize the results of applying the Rasch model to assess the quality of the data. Three copies of the table are provided to facilitate “troubleshooting,” as the results may suggest dropping problematic items or discrepant cases from the analysis, in which case the analysis may be re-run multiple times to evaluate these decisions. You may indicate which items you included at the top of each table, i.e. “Trial 1: all items included,” or “Trial 2: dropped HUNGRY.” Trial 1: all items included Output Acceptable range Problematic items Interpretation Decision/action/notes Infit 0.7-1.3 - Outfit <2 Worried 3.405 Found 16 cases have inconsistency pattern which is equal 0.13% (less than threshold 0.25%) could be ignored Residual correlation < |0.4| for a pair of items - Rasch reliability >0.7 (8-item scale) or >0.6 (7-item scale) Trial 2: Output Acceptable range Problematic items Interpretation Decision/action/notes Infit 0.7-1.3 - Outfit <2 - Residual correlation < |0.4| for a pair of items - Rasch reliability >0.7 (8-item scale) or >0.6 (7-item scale)